Although the random forest classification procedure works well in datasets with many features, when the number of features is huge and the percentage of truly informative features is small, such as with DNA microarray data, its performance tends to decline significantly. In such instances, the procedure can be improved by reducing the contribution of trees whose nodes are populated by non-informative features. To some extent, this can be achieved by prefiltering, but we propose a novel, yet simple, adjustment that has demonstrably superior performance: choose the eligible subsets at each node by weighted random sampling instead of simple random sampling, with the weights tilted in favor of the informative features. This results in an 'enriched random forest'. We illustrate the superior performance of this procedure in several actual microarray datasets.

Probe-level data from Affymetrix GeneChips can be summarized in many ways to produce probe-set level gene expression measures (GEMs). Disturbingly, the different approaches not only generate quite different measures but they could also yield very different analysis results. Here, we explore the question of how much the analysis results really do differ, first at the gene level, then at the biological process level. We demonstrate that, even though the gene level results may not necessarily match each other particularly well, as long as there is reasonably strong differentiation between the groups in the data, the various GEMs do in fact produce results that are similar to one another at the biological process level. Not only that the results are biologically relevant. As the extent of differentiation drops, the degree of concurrence weakens, although the biological relevance of findings at the biological process level may yet remain.

MOTIVATION: DNA microarray technology typically generates many measurements of which only a relatively small subset is informative for the interpretation of the experiment. To avoid false positive results, it is therefore critical to select the informative genes from the large noisy data before the actual analysis. Most currently available filtering techniques are supervised and therefore suffer from a potential risk of overfitting. The unsupervised filtering techniques, on the other hand, are either not very efficient or too stringent as they may mix up signal with noise. We propose to use the multiple probes measuring the same target mRNA as repeated measures to quantify the signal-to-noise ratio of that specific probe set. A Bayesian factor analysis with specifically chosen prior settings, which models this probe level information, is providing an objective feature filtering technique, named informative/non-informative calls (I/NI calls). RESULTS: Based on 30 real-life data sets (including various human, rat, mice and Arabidopsis studies) and a spiked-in data set, it is shown that I/NI calls is highly effective, with exclusion rates ranging from 70% to 99%. Consequently, it offers a critical solution to the curse of high-dimensionality in the analysis of microarray data. AVAILABILITY: This filtering approach is publicly available as a function implemented in the R package FARMS (www.bioinf.jku.at/software/farms/farms.html).